Developmental Biology 291 (2006) 342–355 www.elsevier.com/locate/ydbio Genomes & Developmental Control Gene expression changes at metamorphosis induced by thyroid hormone in Xenopus laevis tadpoles Biswajit Das a, Liquan Cai a, Mark G. Carter b, Yu-Lan Piao b, Alexei A. Sharov b, ⁎ Minoru S.H. Ko b, Donald D. Brown a, a Department of Embryology, Carnegie Institution of Washington, 3520 San Martin Drive, Baltimore, MD 21218, USA b Developmental Genomics and Aging Section, Laboratory of Genetics, National Institute on Aging, NIH, Baltimore, MD 21224, USA Received for publication 30 September 2005; revised 8 December 2005; accepted 14 December 2005 Available online 3 February 2006 Abstract Thyroid hormone (TH) controlled gene expression profiles have been studied in the tail, hind limb and brain tissues during TH-induced and spontaneous Xenopus laevis metamorphosis. Amplified cRNA probes mixed with a universal standard were hybridized to a set of 21,807-sense strand 60-mer oligonucleotides on each slide representing the entries in X. laevis UniGene Build 48. Most of the up-regulated genes in hind limb and brain are the same. This reflects in part the fact that the initial response to TH induction in both tissues is cell proliferation. A large number of up-regulated genes in the limb and brain programs encode common components of the cell cycle, DNA and RNA metabolism, transcription and translation. Notch is one of the few genes that is differentially expressed exclusively in the brain in the first 48 h of TH induction studied in these experiments. The TH-induced gene expression changes in the tail are different from the limb and brain programs. Distinct muscle and fibroblast programs were identified in the tail. Dying muscle fibers in tail (marked by active caspase-3) up-regulate a group of genes that include proteolytic enzymes. At the climax of metamorphosis, tail muscle down-regulates more than half of the genes that encode the glycolytic enzymes in the cytoplasm and the tricarboxylic acid pathway and all five complexes of the electron transport system in mitochondria. These changes in gene expression precede the activation of caspase-3. Some of these same energy metabolism-related genes are up-regulated in the limb and brain programs by TH. A prominent feature of the tail fibroblasts is the down-regulation of several collagen and other extra cellular matrix genes and the up-regulation of hydrolytic enzymes that are responsible for dissolving the notochord and resorbing the tail. © 2005 Elsevier Inc. All rights reserved. Keywords: Thyroid hormone; Metamorphosis; Tadpole; Xenopus laevis; Tail resorption; Limb growth; Brain ventricle proliferation; Mitochondrial electron transport chain; Cell cycle; Transcriptional regulation Introduction 1986). These receptors function as transcription factors. Therefore, changes in gene expression are presumed to be During amphibian metamorphosis, thyroid hormone (TH) at the heart of the remarkable developmental changes that controls developmental changes that range from complete occur during amphibian metamorphosis. The control of so organ growth such as limb development to cell death in the many diverse developmental programs by a single small gills and tail (Dodd and Dodd, 1976). Many tadpole organs molecule makes it a tractable developmental system to study are induced by TH to remodel including the intestine using molecular tools. We have shown previously that the (McAvoy, 1977), pancreas (Dodd and Dodd, 1976), liver thyroid receptors are essential for many if not all of these (Cohen, 1970), and brain (Kollros, 1981). TH functions by developmental programs (Schreiber et al., 2001). binding to thyroid hormone receptors that belong to the More than 10 years ago, we analyzed TH-induced gene nuclear receptor family (Sap et al., 1986; Weinberger et al., expression changes in tail (Brown et al., 1996; Wang and Brown, 1993), limb (Buckbinder and Brown, 1992), intestine ⁎ Corresponding author. (Shi and Brown, 1993) and cultured cells (Kanamori and E-mail address: [email protected] (D.D. Brown). Brown, 1993) by a subtractive hybridization method (Wang 0012-1606/$ - see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2005.12.032 B. Das et al. / Developmental Biology 291 (2006) 342–355 343 and Brown, 1991). In this paper, we have examined the TH- AMADID #012454. The 5445 oligonucleotides that hybridized with intensity induced global gene expression changes in three different below 2.3 in these experiments have been changed in a second version of X. laevis arrays (AMADID #013214). developmental programs (tail, hind limb and brain) using oligonucleotide microarrays designed to contain representative RNA sample collection, probe preparation, and in situ hybridization sequences from all of the X. laevis clusters (Build 48) in the NCBI UniGene database. This global approach has identified NF54 pre-metamorphic tadpoles (Nieuwkoop and Faber, 1956) were many new TH-regulated genes. Clustering of functionally treated in their rearing water with 100 nM T3 for 24 h and 48 h. Tail, hind related genes that are differentially expressed reveal insights limb and brain tissues were isolated (Fig. 1A). The tadpoles collected for into the biological changes induced by TH at metamorphosis. brain and hind limb but not for the tail samples had been pretreated in 1 mM methimazole for 1 week to reduce the endogenous levels of TH so that the response to TH is solely from the externally added hormone (Cai and Materials and methods Brown, 2004). An additional TH-induced early time point (14 h) was analyzed for the limbs to identify possible direct response genes. Tails at the Design of oligonucleotide microarray using the UniGene database for climax of spontaneous metamorphosis (NF62) were also collected. Total Xenopus laevis cDNA sequences RNA was isolated from the dissected tissue samples using the TRIZOL (Invitrogen) method according to manufacturer's protocol. Three separate Agilent Technologies (Palo Alto, CA) prepared the micro array slides groups of animals were treated identically with TH, and tissue samples were using sequences from X. laevis UniGene Build 48 (February 2004). Each slide taken for each time point. Six tadpoles were sacrificed for each tail sample has a capacity for 22,543 oligonucleotides and included 21807 entries and 12 tadpoles for each of the limb and brain samples. The brain and limb representing 21654 UniGene clusters. Each oligonucleotide is in the sense samples were collected from the same tadpoles. cDNA was prepared from direction of the mRNA and 60 nucleotides in length. We added a second the total RNA of each sample. Then, cRNA labeled with Cy3 CTP (Perkin oligonucleotide for a set of 153 genes to serve as internal control. The Elmer Cat #NEL 580) was prepared using a linear amplification and duplicated genes include all of the up- and down-regulated genes that had labeling method (Agilent Kit #5184-3523). This experimental cRNA probe been identified in the TH-induced tail subtractive hybridization studies was mixed with Cy5 CTP (Perkin Elmer, Cat #NEL 581) labeled universal (Brown et al., 1996), cell cycle-related genes, and genes involved in major standard cRNA. The standard cRNA probe was prepared from total RNA signal transduction pathways. UniGene has many duplicate entries because the that had been isolated from whole tadpoles at NF stages 50, 52, 54, 56, 58, X. laevis genome is pseudotetraploid. In addition, Agilent includes 736 60, 62, 64, 66 and juvenile frogs and then combined in equal amounts. We oligonucleotides on each slide as positive and negative hybridization controls. prepared enough universal standard RNA so that it can be used for future These X. laevis arrays are available from Agilent Technologies. The original metamorphosis array experiments permitting a comparison of data from design which was used in these experiments has the reference number different time-series and tissues. In situ hybridization on tissue sections used DIG-labeled RNA probes (Cai and Brown, 2004). Results Statistical analysis, filtering and GO mapping Data from all of the replicates were subjected to correlation matrix analysis, and replicates with a correlation coefficient less than 0.95 were disregarded in further analysis (we removed one sample each from Tail 24 h T3 treatment group, tail NF62 group and one sample from brain 48 h T3 treatment group). We also disregarded the data from 5448 spots that hybridized with mean log-intensity values of less than 2.3 for the Cy5 labeled universal standard cRNA. Differential expression differences in pairwise comparisons used a False Discovery Rate (FDR) method. Gene expression changes are considered statistically significant when their FDR is b5% using an ANOVA-FDR test (Benjamini and Hochberg, 1995; Sharov et al., 2005) (http://lgsun.grc.nia.nih.gov/ANOVA/). All data have been submitted to GEO (NCBI) database (GEO Accession for tail series GSE3405, for limb series GSE3404 and brain series GSE3402). The ANOVA output including the lists of pairwise comparisons of treatment groups, hierarchical clustering, Principal component analyses within each tissue can be accessed at these addresses: Tail data: http://lgsun.grc. nia.nih.gov/ANOVA/output/DBrown-TailArray.html;limb Fig. 1. Overview of the three programs. (A) NF53 tadpole showing the organs in red that was selected to make probes for hybridization. (B) Graph showing the data: http://lgsun.grc.nia.nih.gov/ANOVA/output/DBrown- number of up- and down-regulated genes that